ARTICLE https://doi.org/10.1038/s41467-019-09913-4 OPEN Lipopolysaccharide inhalation recruits and subsets to the alveolar airspace

Laura Jardine 1,2,4, Sarah Wiscombe1,2,4, Gary Reynolds1,2, David McDonald1, Andrew Fuller1, Kile Green1, Andrew Filby1, Ian Forrest2,Marie-HeleneRuchaud-Sparagano1, Jonathan Scott1, Matthew Collin1,2, Muzlifah Haniffa 1,3,4 & A. John Simpson1,2,4

Mononuclear phagocytes (MPs) including monocytes, and dendritic cells

1234567890():,; (DCs) are critical innate immune effectors and initiators of the adaptive immune response. MPs are present in the alveolar airspace at steady state, however little is known about DC recruitment in acute pulmonary inflammation. Here we use lipopolysaccharide inhalation to induce acute inflammation in healthy volunteers and examine the impact on bronchoalveolar lavage fluid and blood MP repertoire. Classical monocytes and two DC subsets (DC2/3 and DC5) are expanded in bronchoalveolar lavage fluid 8 h after lipopolysaccharide inhalation. Surface phenotyping, gene expression profiling and parallel analysis of blood indicate recruited DCs are blood-derived. Recruited monocytes and DCs rapidly adopt typical airspace-resident MP gene expression profiles. Following lipopolysaccharide inhalation, alveolar macrophages strongly up-regulate cytokines for MP recruitment. Our study defines the characteristics of human DCs and monocytes recruited into bronchoalveolar space immediately following localised acute inflammatory stimulus in vivo.

1 Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. 2 Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne NE7 7DN, UK. 3 Department of Dermatology and NIHR Newcastle Biomedical Research Centre, Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne NE2 4LP, UK. 4These authors contributed equally: Laura Jardine, Sarah Wiscombe, Muzlifah Haniffa, A. John Simpson. Correspondence and requests for materials should be addressed to L.J. (email: [email protected]) or to M.H. (email: [email protected])

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

he alveoli of the lung present a large but fragile surface events in human lung inflammation. Eight MP subsets can be area to the environment. Maintaining the integrity of identified in steady state BAL following sterile saline inhalation Tthe alveolar-capillary membrane is critical to effective gas (SS-BAL), in line with subsets described in blood. Monocytes and exchange. Immune regulation at this interface must control two myeloid DC subsets (DC2/3 and DC5) are recruited as early infection and limit immunopathology. Mononuclear phagocytes as 8 h following LPS inhalation (LPS-BAL) and rapidly adopt (MPs), comprising monocytes, macrophages, and dendritic cells gene expression profiles characteristic of airspace MP residence. (DCs) have a critical role at any environmental interface as innate The cytokine and chemokine profile of BAL implicates AMs as immune effectors equipped to shape the adaptive immune the likely instigator of blood and DC recruitment into response through antigen presentation, co-stimulation, and the alveolar airspace during acute inflammation. cytokine production. Leukocytes from the alveolar airspace can be readily isolated by bronchoscopy and bronchoalveolar lavage (BAL). BAL has been Results less extensively characterized than lung tissue but presents a Accumulation of alveolar neutrophils, monocytes, and DCs. number of advantages as a window to the lung’s immune system: The steady state MP repertoire of BAL was defined by flow BAL yields a cell suspension free from contaminating blood cytometry in healthy individuals 8 h after inhalation of isotonic leukocytes with minimal processing requirement thus preserving saline (Fig. 1a, Supplementary Fig. 1). We used a recent surface antigens and native activation status. description of human blood MP diversity based on single cell MPs in any anatomical compartment are a heterogeneous RNA-sequencing as a template for identifying BAL MP popula- group of leukocytes. Establishing the steady-state repertoire of a tions19. As an adaption for BAL analysis, we first used side scatter compartment is crucial to understanding infiltrates seen in and CD45 expression to exclude CD45loSSCmid neutrophils and inflammation. Most tissues contain embryonically-derived mac- identify alveolar macrophages (AM) as CD45+SSChi cells (Sup- rophages with variable contributions from circulating monocytes plementary Fig. 1). BAL CD45+SSClo cells were then comparable depending on how available the tissue niche is1,2. Monocyte- to peripheral blood mononuclear cells (PBMC) (Fig. 1b). Per- derived cells are observed adopting a spectrum of ipheral blood was sourced from healthy controls (HC) that had or DC features depending on the tissue. While steady state not received inhalation challenge. CD45+SSClo cells that monocyte-derived DCs are observed in mouse skin and gut3–5, expressed HLA-DR but not lineage markers (CD3, 19, 20, 56) their existence in human tissues has not been convincingly were gated into CD14++CD16−, CD14++CD16+, and CD14+ established. DCs are broadly divided into plasmacytoid DCs CD16++ fractions analogous to blood classical, intermediate and (pDC) that characteristically produce IFN-α and conventional non-classical monocytes, respectively. The CD14−CD16− frac- DCs (cDC) that effectively stimulate T cell proliferation6. Two tion containing DCs was first probed for the rare Axl+Siglec6+ subsets of cDCs with homology across species have been clearly population (DC5) as its varied expression of CD123 and CD11c identified. cDC1 expresses CD141, CLEC9A, and XCR1 in would otherwise place it within pDC and cDC gates. After DC5 humans and is adept at cross-presenting antigen7–11. cDC2 exclusion, CD123+CD11clo cells identified pDC and further expresses CD1c in humans and is important for activation of CD4 segregation of the CD11c+ cells using a combination of BTLA T cells, induction of regulatory T cells and activation of Th2 and and CD1c identified DC1 and DC2/3. While DC1 is most typi- Th17 responses12,13. cally defined by its expression of CD141 or CLEC9A, immune A number of recent studies have capitalized on multi- gene expression by cells sorted from the BTLAhi gate expressed parameter flow cytometry to define subsets of MPs across the expected gene profile of DC1, confirming the validity of this human lung compartments14–18, but differences in approach have approach (Supplementary Fig. 1). We did detect a population led to continued debate about whether rare DC subsets (pDC and of CD11c+CD1c−BTLA− cells within our DC gate, which are cDC1) exist in BAL or have been inadvertently excluded during likely to correspond to DC4 described in Villani et al., but can analysis. As blood is a source of tissue-recruited leukocytes, a only refer to these as CD11c+CD1c− cells without further logical approach would use blood MP definitions to classify tissue characterization. MPs. Recent insights from single cell RNA sequencing have The dominant MP subset in SS-BAL was the AM (Table 1). revealed additional complexity in our understanding of blood Comparing SS-BAL CD45+SSClo cells with PBMC, MPs were MPs, including the presence of previously undiscovered DC richer in SS-BAL (Table 1). Of MPs, the most abundant subset subsets, and heterogeneity within existing subsets19. Briefly, this was a CD14++CD16+ population (Table 1; Fig. 1d). CD14++ confirmed the presence of cDC1 (DC1) and pDC (pDC or DC6). CD16− cells, analogous to classical monocytes in blood, were It revealed two subdivisions within cDC2 (DC2, DC3) and comparatively rare in SS-BAL and CD14+CD16++ cells, identified additional DCs subsets: Axl+Siglec−6+ DCs (AS DC analogous to non-classical blood monocytes, were virtually or DC5) and CD1c-CD141− DCs (DC4). To date, this revised absent. All DC subsets, especially CD1c-expressing DC1 and classification has not been tested in non-lymphoid tissue. DC2, were enriched in SS-BAL relative to blood. As our understanding of the BAL MP repertoire in steady state Following LPS inhalation, the greatest leukocyte expansion was develops, it becomes tangible to address the question of what in CD14++CD16− MPs (400-fold difference in mean concentra- happens during inflammation. In mice, inflammatory macrophages tion between SS-BAL and LPS-BAL), followed by neutrophils and DCs have been described in numerous infection and sterile and DCs (Fig. 1c). Amongst DCs, the concentrations of CD1c+ inflammation models20–25. Monocytes are thought to be the source DCs (DC2/3) and DC5 were selectively increased (Fig. 1c). of inflammatory DCs, based on studies in CCR2 and Flt3-deficient To examine the impact of acute lung inflammation on animals and adoptive transfer of monocytes20–24.Distinctinflam- peripheral blood MP populations, their concentrations were matory macrophages and DCs have also been identified in human tracked at 2-h intervals following inhalation of LPS or saline chronic inflammatory exudates26.Theseinflammatory DCs are (Fig. 1e). Neutrophils, which were abundant in LPS-BAL and proposed to arise from monocytes based on transcriptional simi- pDCs, which were not enriched in LPS-BAL, were tracked for larity to in vitro monocyte-derived DCs26.Earlytime-pointsof comparison with monocytes and DCs. Following saline inhala- inflammation have not been adequately explored. tion, no differences in leukocyte concentrations occurred Here, we use inhaled lipopolysaccharide (LPS) as an experi- compared with baseline. Following LPS inhalation, blood mental inflammatory stimulus to reproducibly study the earliest neutrophil concentration rose progressively, but pDC and

2 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

a c Neut AM CD14++CD16– Inhalation BAL 100 **** 100 ns 100 ** Saline 80 80 80

LPS 60 60 60 0 2 4 6 8 hours 40 40 40 20 20 20

0 0 0 CD14++CD16+ DC2/3 CD14+CD16++ 20 ** 8 *** 2.0 ns

15 6 1.5

cells/ L BAL cells/ L SS-BAL 6 10 4 1.0 LPS-BAL b HC blood SS-BAL LPS-BAL × 10 2 0.5 250K 250K 250K 5

200K 200K 200K 0 0 0 150K 150K 150K pDC CD1c- DC DC1 DC5 100K 100K 100K 2.0 ns 2.5 ns 1.0 ns 1.0 **

SSC-A 50K 50K 50K 2.0 1.5 0 0 0 1.5 2 3 4 5 2 3 4 5 2 3 4 5 CD45 0 10 10 10 10 010 10 10 10 010 10 10 10 1.0 0.5 0.5 1.0 5 5 10 10 105 0.5 0.5 104 104 104 0 0 0 0 103 103 103 102 102 102 d HC blood SS-BAL LPS-BAL HLA-DR 0 0 0

0 103 104 105 0 103 104 105 0 103 104 105 lin 5 10 105 105 4 10 104 104 3 10 103 103 0 0 0 % of non-AM MPs CD14

0 103 104 105 0 103 104 105 0 103 104 105 e Blood post saline Blood post LPS CD16 inhalation inhalation 5 5 105 10 10 12 ×109 ns 12 *** 4 4 104 10 10 8 8 Neutrophil 103 103 3 10 4 4 0 0

Axl 2 10 0 0 0 0246 0246 0 103 104105 0 103 104105 0 103 104 105 8 Siglec-6 10 ×10 ns 10 ns 8 8 105 105 105 6 6 Monocyte 4 4 4 (total) 104 10 104 2 2 3 3 103 10 10 0 0 0246 0246 2 10 0 0 4 CD123 1.5 1.5 0 ×10 ns * DC2/3 2 3 4 5 2 3 4 5 2 3 4 5 1.0 1.0 CD11c 010 10 10 10 010 10 10 10 0 10 10 10 10 5 0.5 0.5 10 105 105 104 0 0 104 104 Cells/ L blood Cells/ L 0246 0246 103 103 103 2.0 ×102 ns 2.0 ns 2 10 1.5 1.5 0

BTLA DC5 0 0 1.0 1.0 0.5 0.5 0102 103 104 105 0 103 104 105 0 103 104 105 CD1c 0 0 0246 0246 Neut CD14++CD16– DC1 1.5 4 ×10 ns 1.5 ns CD14++CD16+ DC2/3 AM 1.0 1.0 pDC CD14+CD16++ DC5 0.5 0.5 pDC 0 0 CD1c- DC 0246 0246 Time (hours) following inhalation monocyte concentrations remained static. Blood DC2/3 concen- changes of tissue adaptation in vivo following acute inflammatory tration fell significantly within 2 h of LPS inhalation. challenge. The surface antigen phenotype of LPS-BAL CD14++ CD16− MPs by flow cytometry resembled that of blood mono- Alveolar CD14++CD16− MPs are recruited blood monocytes. cytes. LPS-BAL CD14++CD16− cells expressed genes char- As CD14++CD16− MPs were particularly enriched in LPS-BAL, acteristic of classical monocytes including CD14, CCR2 and SELL we focused on this population to interrogate the molecular but did not express FCGR3A/B (CD16), CD79b, and CX3CR1,

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

Fig. 1 Neutrophils and mononuclear phagocytes are expanded in the alveolar airspace following LPS inhalation. a Schematic overview of study design. Solid arrows denote LPS inhalation (red) or saline inhalation control (blue). Black arrows denote blood sampling. Dashed arrows denote BAL. b Flow cytometry of leukocyte preparations from SS-BAL, LPS-BAL, and HC blood. The CD45 versus SSC plot was used to define CD45+SSChi AM, CD45loSSCmid neutrophils and CD45+SSClo mononuclear cells (see also Supplementary Fig. 1). Monocyte/macrophages and DCs were negative for lineage markers CD3, CD19, CD20, and CD56 and expressed HLA-DR. Monocyte/macrophages were divided into CD14++CD16−, CD14++CD16+, and CD14-CD16++ populations analogous to blood classical, intermediate and non-classical monocytes. DCs within the CD14-CD16− gate were divided into subsets: Axl+ Siglec6+ DC5s, CD11cloCD123+ pDCs, CD1c+BTLAlo-mid DC2/3s, and BTLAhi DC1s. Plots are representative of n = 9 SS BAL and n = 10 LPS BAL. c Concentrations of neutrophil, monocyte/macrophage and DC subsets in SS-BAL and LPS BAL. Bars represent mean and lines SEM. p-values from unpaired t-tests of SS versus LPS are shown: “ns” p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. d Monocyte/macrophage and DC frequency in HC blood, SS-BAL and LPS-BAL as a proportion of SSClo MHC class II-expressing cells (not-including CD1c- DCs). e Concentration of selected leukocyte populations in peripheral blood at 2-h intervals following inhalation of saline (blue line) or LPS (red line). Data points show mean ± SEM for 3–5 participants. p-values from one-way ANOVA are shown. p-value representation is described in c

Table 1 Frequency of MP subsets in BAL and blood

MP subset AM CD14++ CD14++ CD14+ DC1 DC2/3 CD1c− DC5 pDC CD16− CD16+ CD16++ Antigen CD45 ++ + + + + + + + expression SSC hi lo lo lo lo lo lo lo lo by flow Lineage −− − − − − − − − cytometry HLA-DR ++ + + + + + + + CD14 + ++ ++ + −−−−− CD16 + − + ++ −−−−− CD11c +++-/+ − CD123 −−−−/++ BTLA ++ −/+ − ++ CD1c +/− + − -/+ − Axl +/− +/− +/− + − Siglec 6 −−−+ − % of l HC whole blood 0 8.01 (2.04) 0.47 (0.22) 1.29 (0.85) 0.02 (0.01) 0.30 (0.1) 0.07 (0.07) 0.02 (0.02) 0.23 (0.05) eukocytes 75.0 (7.94) 6.41 (2.29) 13.8 (6.11) 0.23 (0.22) 1.07 (0.75) 0.77 (0.72) 0.22 (0.20) 1.93 (0.93) mean (SD) SS BAL 60.1 (14.3) 1.06 (0.14) 2.38 (0.95) 0.64 (0.56) 0.12 (0.10) 0.82 (0.35) 0.69 (0.72) 0.04 (0.02) 0.13 (0.07) % of MPs 2.92 (1.88) 43.7 (8.00) 11.4 (7.31) 2.41 (1.91) 16.1 (4.24) 12.3 (13.9) 1.06 (0.36) 2.57 (1.53) mean (SD) LPS BAL 23.1 (13.5) 18.7 (9.13) 4.51 (3.23) 0.23 (0.16) 0.09 (0.04) 1.63 (0.61) 0.2 (0.11) 0.14 (0.05) 0.22 (0.16) 64.3 (17.8) 18.2 (16.6) 0.91 (0.67) 0.33 (0.15) 6.12 (2.59) 0.69 (0.37) 0.45 (0.11) 0.76 (0.49)

MP subsets present in BAL. MP subsets are defined by the flow cytometry gates in Fig. 1b. Expression of defining antigens is indicated as “++” bright expression, “+” positive expression, “+/−” low-level expression or “−” no expression. Where no symbol is given, expression was not measured on that subset. Frequencies of each MP subset in whole blood, SS BAL and LPS BAL are given. In standard type, frequencies are given as a percentage of total leukocytes (defined as CD45+ cells). In italic type, frequencies are given as a percentage of MPs (defined as CD45+SSClo, CD3,19,20,56− HLA-DR+ cells) which are characteristic of non-classical monocytes27 (Fig. 2a). that CD14++CD16− MPs are recruited blood monocytes with However, unsupervised transcriptome analysis revealed distinct maturation and adaptation to the alveolar environment. changes as a consequence of recruitment into the alveolar Comparing CD14++CD16− MPs in SS-BAL and LPS-BAL, 98 airspace. genes (21%) were differentially expressed (Fig. 2d, Supplementary Using the NanoString V2 579-gene panel, we Dataset 1). Steady state (SS-BAL) was associated with higher compared the expression profile of HC blood monocyte subsets expression genes involved in antigen presentation (CD1A, HLA- with SS-BAL and LPS-BAL monocyte-macrophages. By principal DPA1, HLA-DMB, HLA-DOB, CD1D), pathogen clearance component analysis, the first principal component (PC1) (CD209, MRC1), control of inflammatory signaling (PTPN22), segregated cells by tissue compartment (blood versus BAL) and and the ability to activate the allergic responses (FCEFR1A). the second (PC2) separated monocyte-macrophages from resi- Inflammation (LPS-BAL) saw activation of LPS response genes dent alveolar macrophages (Fig. 2b). PC3 separated the CD14+ (IL1A, IL1B, IRAK2, TRAF3, and others) retention of monocyte- +CD16− cells from SS-BAL and LPS-BAL. The distance between associated genes (S100A8/9, SELL), and marked expression of BAL CD14++CD16− MPs and monocytes, even by principal chemokine genes (CCL2/3/4/5/7, CXCL10, and IL8). This suggests components 2 and 3, suggested adoption of a unique gene that in quiescence, CD14++CD16− MPs have the capacity to expression profile upon entry to the airspace. This was explored present antigen, possibly helping to maintain tolerance, while in further by examination of differentially expressed genes (DEGs). inflammation, they are primed to modulate immune functions We performed pairwise comparisons of DEGs of CD14++ through chemokine and interleukin production. CD16− cells between tissue compartments (HC blood vs. BAL) and between inflammatory states (SS-BAL vs. LPS-BAL). There Blood CD1c+ DC are recruited into the alveolar space. We used was a core signature of 49 DEGs (11% of the 457 genes expressed a panel of antigens showing divergent expression on tissue and in these cell types) distinguishing SS-BAL and LPS-BAL CD14++ blood CD1c-expressing DCs (DC2/3) to evaluate the flow cyto- CD16− MPs from HC blood classical monocytes (Fig. 2c, metry phenotype of CD1c+ DCs found in SS-BAL, LPS-BAL and Supplementary Dataset 1). The 36 up-regulated included genes HC blood14,26,28 (Fig. 3a). Blood DC2/3 were CD11blo, Axl−, encoding phagocytic receptors (MRC1, FCGR3A/B, MSR1), CD1a−, and CD206− while SS-BAL DC2/3 expressed all four of immunoregulatory proteins (CD274), innate immune effectors these antigens, confirming that these were genuine tissue DCs and (C1QA/B), chemokines capable of monocyte and lymphocyte not blood contaminants. However, the expression of these anti- recruitment (CXCL10), and some genes associated with mature gens on LPS-BAL DC2/3 paralleled that seen in blood DC2/3. DC function (CD80, LAMP3). This supported the interpretation Significant blood contamination of LPS-BAL was excluded by

4 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

a b CD14 CD16 8000 20,000 BAL 6000 15,000 SS CD14++CD16– SS CD14++CD16+ 10,000 4000 PC1 (40%) SS Alveolar macrophage 2000 5000 LPS CD14++CD16– LPS Alveolar macrophage 0 0 Blood 800 CCR2 800 CD79b Classical monocyte 600 600 PC2 (19%) Intermediate monocyte 400 400 Non-classical monocyte

200 200

Gene expression (count) 0 0 2000 SELL 2500 CX3CR1 PC2 (19%) 2000 1500 1500 1000 1000 500 500 PC3 (8%) 0 0 d BAL CD14++CD16– SS LPS

c NFKBIA SS-BAL vs blood LPS-BAL vs blood S100A9 CCL2 53 73 CXCL10 IL8 IL1B CCR6 CFP 49 IL1A CD1D CCL3 NFATC2 S100A8 CCL20 CD244 S100A8 FCER1A S100A9 CCL7 TNFAIP3 CD79B LCP2 CD9 TNFRSF8 CCL5 ENTPD1 IL1RN SOCS1 LILRA3 TNFRSF4 IL11RA GBP1 CCL24 CCL4 LILRA5 ICAM1 HLA-DQA1 CD22 LILRB2 CD1A IRAK2 SRC CD59 TNFAIP6 NLRP3 CEACAM8 CCL3 CLEC4E NFKB2 IRAK1 CCRL2 GBP5 IL1A IRAK2 BAL CD14++CD16- Blood classical IRF1 TNFSF8 SS LPS monocyte IKBKE BCL3 MRC1 TRAF3 CCL2 TRAF1 LITAF JAK3 SOCS3 SELL CXCL10 CFP CDKN1A RELB CLEC5A SMAD3 MSR1 TNFAIP6 GBP1 CCL7 C1QB CD82 CXCL9 CCL5 FCGR3A/B CD22 IFIT2 CD45RA PLAUR TNFSF15 GPR183 CSF1 SERPING1 TNFRSF4 C1QA BATF CXCL11 TNFRSF8 TFRC CCL23 CD80 IL18R1 PLAU CD6 BATF3 HFE CISH SLC2A1 IDO1 TRAF4 IL7R CD209 LAMP3 MAP4K1 CD276 PTPN22 CD274 Relative CD1A SLAMF7 ENTPD1 CCL18 gene NFATC2 CD244 expression CCR6 CD36 CD9 FCER1A Relative gene 0 6 12 FCGR2B 0 6 12 expression MRC1 counting erythrocytes: 16.7 ×106 ± 28.4 ×106 in SS-BAL and due to blood contaminants (DC2/3 frequency in peripheral 33.3 ×106 ± 30.8 ×106 in LPS-BAL (mean ± SD; p = 0.04 by blood is approximately 10 cells μl−129). unpaired t-test). This amounted to a 6 μl blood leak, permitting While the surface antigen phenotype of LPS-BAL DC2/3 by approximately 0.4% of the 105 DC2/3 found in LPS-BAL to be flow cytometry supported their rapid recruitment from blood

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

Fig. 2 CD14++CD16− monocyte-macrophage cells in LPS-BAL are recruited blood monocytes with transcriptional adaptations. a Expression of monocyte- subset discriminating genes in CD14++CD16− cells from SS-BAL (blue square) and LPS-BAL (red square) compared with HC blood classical (filled green square), intermediate (divided green square) and non-classical (open green square) monocytes. Gene expression was quantified by NanoString array. The left-hand column contains genes with higher expression in classical than intermediate or non-classical monocytes. The right-hand column contains genes with higher expression in intermediate and non-classical than classical monocytes. Bars represent mean ± SEM. b Principal component analysis of immune gene expression (579-gene NanosString array) by BAL monocyte/macrophages and blood monocyte subsets. c Analysis of differentially expressed genes (DEGs) in BAL CD14++CD16− cells and HC blood monocytes. Comparisons were made by unpaired t-test with p < 0.05 and >3-fold difference in mean expression. Venn diagram shows DEGs in CD14++CD16− MPs from SS-BAL, LPS-BAL and HC blood. Circles represent DEGs between SS BAL and blood (cyan; 101 genes, 22%) and between LPS-BAL and blood (yellow; 124 genes, 27%). The overlapping circles represents DEGs shared between comparisons (49 genes, 11%). The top 10 upregulated genes (standard type) and top 5 downregulated genes (italic type) in BAL relative to blood are listed. Heatmap shows DEGs common to both SS-BAL and LPS-BAL CD14++CD16- cells relative to HC blood classical monocytes. “Relative gene expression” refers to log2 transformed normalized gene expression count data. Genes with >10-fold difference in mean expression are shown. Genes discussed in text are colored burgundy. d Heatmap showing DEGs between SS-BAL and LPS-BAL CD14++CD16− cells. Comparisons were made by unpaired t-test with p < 0.05. Genes with >5-fold difference in expression are shown. Genes discussed in the text are colored burgundy

DC2/3, consideration was also given to the possibility that they and DC3 (BST1, CD14, CD163, S100A8/9)19 (Fig. 4a) and with may have differentiated from recruited blood monocytes. The blood CD1c+ DC subsets segregated by CD5 expression30. latter possibility was considered less likely within the eight-hour In BAL, BTLA expression discerned two subsets with the mean time window studied. Microarray analysis in human chronic fluorescence intensity for BTLA+ cells at 1 ×103 (corresponding inflammatory exudates aligned inflammatory CD1c-expressing to DC2) and BTLA− cells at 100 (corresponding to DC3) DCs with in vitro monocyte-derived DCs26. From a list of (Fig. 4b). The ratio of DC2:DC3 in HC blood was 40:60. In conserved genes defining human and mouse monocytes and contrast, SS-BAL showed marked skewing towards DC3 (20:80). macrophages compared to DCs28, 25 genes were evaluable on the In keeping with our previous findings suggesting that LPS-BAL Nanostring Immunology V2 panel. Hierarchical clustering the CD1c+ DCs are recruited from circulating DCs, the DC2:DC3 expression of these genes located LPS-BAL DC2/3 closest to SS- ratio in LPS-BAL approximated that of blood. BAL DC2/3 and HC blood DC2/3 and remote from HC blood The genes differentially expressed between HC blood DC2 and monocytes, BAL macrophages and in vitro monocyte-derived DC3 were not recapitulated between LPS-BAL DC2 and DC3 DCs (Fig. 3b). Furthermore, LPS-BAL CD1c-expressing MPs (Fig. 4c). The convergence in gene expression between LPS-BAL were effective stimulators of allogeneic T cell proliferation, in CD1c+ DC subsets following recruitment to BAL may be contrast to LPS-BAL CD14++CD16− MPs and AMs, further accounted for by inflammation (SLAMF7, TNFRSF9), as well as supporting their origin from circulating CD1c-expressing blood change of compartment (CXCL9, FCGR3A/B, S100A8/9). DCs (Fig. 3c). Both blood CD1c+ subsets effectively stimulate allogeneic T We next explored the functional adaptations to airspace cell production19,30 but have distinct capacity for inducing residence of recruited blood DCs by comparing the immune cytokine production by CD4+ T cells, with CD5+CD1c+ DCs gene expression profile of blood DC2/3 with their SS-BAL and (representing DC2 or BTLA+ DC) inducing IL-10, IL-22, IL-4, LPS-BAL counterparts (Fig. 3d, Supplementary Dataset 1). There and IL-17 production) and CD5loCD1c+ DCs (representing DC3 was a core signature of 100 DEGs in SS and LPS BAL DC2/3 or BTLA- DC) inducing IFN-γ production. We therefore compared with HC blood DC2/3 (21% of the 466 genes expressed compared the capacity of LPS-BAL DC2 and DC3 to influence in these cell types): 39 upregulated and 61 downregulated genes. cytokine production by CD4+ T cells. We observed that both Up-regulated genes were required for mature DC function subsets were equally capable of inducing IFN-γ and IL-17 (CD80, CCR7, LAMP3), involved in immunoregulation (CISH, production without any IL-4 production. The capacity to induce CD274, CD276, TNFRSF11A), monocyte and T cell recruitment IFN-γ production was significantly greater in LPS-BAL than SS- (CCL2, CCL22), and pathogen recognition or scavenging BAL DC2/3. Yet, even SS-BAL DC2 did not induce IL-4 (CLEC5A, MSR1). The genes most down-regulated in BAL production, supporting the concept that CD1c+DCs undergo DC2/3 compared with blood DC2/3 included CD1D, possibly as functional alteration upon tissue entry (Fig. 4d). an adaptation to the lipoprotein-rich alveolar environment, and The Axl+Siglec-6+ DC5 population, while enriched in LPS- the lymphoid-lineage associated genes CD22 and CD244. BAL, remained too small to analyse in functional assays or Relatively few genes were differentially expressed between SS NanoString arrays. DC5 in peripheral blood exhibit a spectrum of and LPS BAL DC2/3 (Fig. 3e, Supplementary Dataset 1). As with CD123 and CD11c expression, with CD123hiCD11clo DC5 LPS CD14++CD16−MPs, LPS DC2/3 expressed LPS response expressing a pDC-like gene signature and CD123loCD11chi genes (IL1B, IL8, IRAK2) and chemokine genes (CCL2,4,5 DC5 expressing a cDC2-like gene signature19. SS-BAL DC5 CXCL10,11) to a greater extent than SS-BAL DC2/3. expressed more CD123, but a greater proportion of LPS-BAL DC5 expressed CD11c (Fig. 4e). This was in keeping with a cDC2 over pDC bias in the inflamed airspace. CD1c+ DC heterogeneity within the alveolar space. The flow cytometry gating strategy used for parallel examination of BAL Alveolar macrophages orchestrate MP recruitment. We hypo- and blood revealed heterogeneity in the CD1c-expressing DC gate thesized that resident AMs orchestrated the recruitment of leu- by BTLA expression (Fig. 1b). In blood, BTLA expression seg- kocytes into the alveolar airspace following LPS inhalation. LPS- regates CD1c-expressing DCs into two subsets with distinct gene BAL AMs demonstrated high expression of chemokine genes expression profiles, with BTLA+ DC expressing lymphocyte including CCL2-4 and CXCL10-11 (Fig. 5a, Supplementary lineage genes including CD5, CD79A/B and CD24 and BTLA− Dataset 1). Compared with SS-BAL AMs, LPS-BAL AMs DC expressing monocyte/macrophage genes such as CD14, expressed 4 to 42-fold higher levels of these chemokine tran- S100A8/9 and F13A1 (Fig. 4a). Gene expression differences scripts. Only CXCL12 (stromal-cell derived factor) was not between blood BTLA+ and BTLA− DC correlate with gene expressed by AMs. Corresponding protein measurements of expression differences between blood DC2 (HLA−class II genes) secreted chemokines confirmed high expression of CCL2–4,

6 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

ab HC blood SS-BAL LPS-BAL HC blood SS-BAL LPS-BAL moDC 100 100 100 100 100 100 80 80 80 80 80 80 60 60 60 60 60 60 40 40 40 40 40 40 20 20 20 20 20 20 0 0 0 0 0 0 3 4 5 3 4 5 3 4 5 HC blood DC2/3 0 103 104 105 0 103104105 0 103104105 0 10 10 10 0 10 10 10 0 10 10 10 CD1c CD1a 100 100 100 100 100 100 80 80 80 80 80 80 60 60 60 60 60 60 SS-BAL 40 40 40 40 40 40 CD14++CD16- 20 20 20 20 20 20 0 0 0 0 0 0 0 103 104 105 0 103104105 0 103104105 0 103 104 105 0 103104 105 0 103104 105 SS-BAL DC2/3 CD11b CD206 100 100 100 100 100 100 80 80 80 80 80 80 60 60 60 60 60 60 40 40 40 40 40 40 20 20 20 20 20 20 0 0 0 0 0 0 LPS-BAL DC2/3 0 103 104 105 0 103104105 0 103104105 0 103 104 105 0 103104 105 0 103104 105 BTLA FcƐR1 100 100 100 100 100 100 80 80 80 80 80 80 60 60 60 60 60 60 40 40 40 40 40 40 20 20 20 20 20 20 0 0 0 0 HC blood 3 4 5 3 4 5 3 4 5 Frequency 3 4 5 3 4 5 3 4 5 Frequency 0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10 classical monocyte Axl CD86 LPS-BAL CD14++CD16- c Gated on CD3+ cells * T cells + LPS-BAL + LPS-BAL + LPS-BAL 40 * SS-BAL AM CD14++CD16- DC2/3 * CD14++CD16- 30 (14) SS-BAL 105 20 CD14++CD16+ 104

3 10 diluted 10 LPS-BAL Alveolar macrophage CD8 0 % CD3 CFSE 0 SS-BAL Alveolar 0 103 104 105 – AM 14 DC2/3 CSFE + allogeneic T cells macrophage

BAL DC2/3 SS-BAL DC2/3 LPS-BAL DC2/3 d SSLPS Blood DC2/3 vs blood DC2/3 vs blood DC2/3

54 81 C1QB C1QA MME CD45RA IRAK2 CD79B SERPING1 MSR1 PPARG SMAD3 CCL4 C5 FCGR3A/B MARCO FYN TNFRSF9 KIT CCL22 CDKN1A CCL19 BATF IL8 CR1 LAMP3 ICOS LILRA5 CD40 IL11RA IL7R PDCD1LG2 CXCL2 CD80 CXCL9 CEACAM8 IL6 CD9 IKZF2 CD83 IDO1 C2 EGR2 100 SRC PLAU CD1A CCL3 CLEC5A CXCL10 CCR7 DUSP4 CD276 CISH TRAF1 e CD274 GBP5 TNFRSF11A BAL DC2/3 CCL2 CXCL11 SS LPS EBI3 SOCS1 IL1A IL1B CCL20 CXCL10 PDGFB CSF1 CCL4 TNFRSF4 CCL2 TNFRSF8 C4BPA IRAK2 TNFSF8 CXCL11 BLNK IL8 LTB4R2 CD2 CCL5 LILRA1 TNFRSF4 CD244 CD22 CD45RA LTB4R C8B CCR2 CD36 MME CFP FN1 MAP4K1 C1QA ICAM3 CD1D SELL FCER1A 0 6 12 Relative gene Relative gene expression expression 06 12

CXCL10, and CXCL12 in LPS-BAL supernatant (Fig. 5b). mRNA Finally, we evaluated the dynamics of pro-inflammatory cytokine profiling of the cognate chemokine receptors revealed their secretion by resident AMs and recruited CD14++CD16− abundance on blood monocytes and DC2/DC3 (Fig. 5c), in monocyte–macrophages isolated from LPS-BAL. Both resident keeping with our findings of their recruitment into BAL following LPS-BAL AMs and recruited LPS-BAL monocyte–macrophages LPS challenge. secreted high levels of pro-inflammatory cytokines upon re-

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

Fig. 3 CD1c+ DCs in LPS-BAL are likely to be tissue-recruited blood CD1c+ DCs. a Expression of surface antigens by flow cytometry on DC2/3 from HC blood (green), SS-BAL (blue) and LPS-BAL (red) relative to isotype control (gray). Antigens predicted to discriminate between blood and tissue CD1c+ DCs were tested. Representative plots from more than three experiments are shown. b Dendrogram showing hierarchical clustering of monocyte/macrophage and DC gene expression by indicated subsets isolated from HC blood and SS/LPS BAL. In vitro monocyte-derived DCs were also included. Clustering used Pearson correlation metric. Monocyte/macrophage and DC identifying genes were taken from McGovern et al.28. The 25 genes available on the NanoString Immunology v2 panel were used. c Proliferation of allogeneic peripheral blood T cells measured by CSFE dilution during 7-day co-culture with or without MP subsets isolated from LPS BAL. Flow cytometry plots are gated on CD3+ T cells from a representative experiment. Summary graph shows 2–3 replicates per subset. Bars show mean ± SEM. Means were compared by one-way ANOVA with Dunnett’s multiple comparison test of DC2/3 against other subsets. *p < 0.05. d Analysis of differentially expressed genes (DEGs) by DC2/3 from SS/LPS-BAL and HC blood. Comparisons were made by unpaired t-test with p < 0.05 and >3-fold difference in mean expression. Venn diagram shows DEGs in DC2/3 from SS-BAL, LPS-BAL, and HC blood. Circles represent DEGs between SS BAL and blood (cyan; 154 genes, 33%) and between LPS-BAL and blood (yellow; 124 genes, 39%). The overlapping circles represents DEGs shared between comparisons (100 genes, 21%). The top 10 upregulated genes (standard type) and top 5 downregulated genes (italic type) in BAL relative to blood are listed. Heatmap shows DEGs common to DC2/3 in both SS-BAL and LPS-BAL relative to HC blood. Genes with >10-fold difference in mean expression are shown. Genes discussed in the text are colored burgundy. e Heatmap showing DEGs between DC2/3 from SS-BAL and LPS-BAL. Genes with >10-fold difference in mean expression are shown. Genes discussed in the text are colored burgundy stimulation with LPS in vitro, suggesting a coordinated partnership immune response as macrophages and DCs differ in capacity for between early resident AM response which is further amplified by migration, phagocytosis, ability to induce adaptive immune recruited CD14++CD16− monocyte–macrophages in mediating responses and in the cytokine/chemokine secretion profile. early acute tissue inflammation (Fig. 5d). Through analysis of immune gene expression we established that recruited alveolar CD14++CD16− cells in LPS-BAL and their counterpart in SS-BAL were distinct from blood monocytes. They Discussion expressed a broad array of genes important for mature macro- The human MP system is a complex repertoire of innate immune phage function including phagocytic receptors, immunor- cells distributed across distinct anatomical compartments that egulatory proteins and cytokines. Despite the LPS activation serve heterogeneous functions in the inflammatory response. signature present in recruited cells, there was commonality in Unraveling this complexity is necessary to understand tissue the gene expression profiles of steady state and inflammatory immune surveillance, immunopathology and inform vaccine CD14++CD16− cells, emphasizing the importance and rapidity design. of the impact of tissue microenvironment on shaping monocyte In keeping with previous reports15,31,32, we identified alveolar fate following extravasation. airspace DC subsets (DC1, DC2/DC3, and pDC) corresponding In addition to the expansion of monocyte–macrophages, to those previously described in peripheral blood. We established we identified several DC subsets expanded in the alveolar that BTLA distinguishes two CD1c+ DC subsets in BAL with airspace. Recruitment of DC from the blood to the gas exchan- equivalent gene expression profiles to DC2/CD5+ DCs and DC3/ ging regions of the previously healthy human lung during CD5lo DCs in blood19,33. Separate CD1c+ subsets have also been early inflammation has never been documented before. This described in skin and arise independently from haematopoietic recruitment was observed for specific DC subsets, including precursors, confirming that these are stable subsets and not CD1c-expressing DC2/3 and Axl+Siglec6+ DC5 but not DC1 and simply variations in activation state33. We showed that BTLA pDCs. Sequential analysis of peripheral blood DC2/3 expression is another reliable marker of DC1 in blood and lung following LPS inhalation detected significant depletion during and verified the identity of sorted BTLAhi cells through their the period of accumulation in BAL. Analysis of surface expression of DC1 defining genes (e.g., XCR1, TLR3, and IRF8). antigens with differential expression between blood and tissue Two other studies14,18 did not identify DC1 in lung interstitial demonstrated that recruited DC had an antigen profile compar- tissue, which may be explained by the experimental design and able with blood, consistent with recruitment of blood DCs and gating strategy used for analysis. We identified a distinct pDC providing evidence against translocation of DCs from the lung population in BAL, consistent with some reports15,17,34 but not interstitium. In the most comprehensive description of human others18. The inconsistent finding of pDC in BAL does not arise inflammatory DCs to date, CD1c-expressing DCs in chronic from their selective vulnerability to cryopreservation17. Although inflammatory exudates were transcriptionally aligned to in vitro Desch et al. did not find a CD123+CD303+pDC population in monocyte-derived DCs26. This fits with numerous observations healthy lung tissue, CD303+ cells were identified in tumor- in mouse that monocyte-derived cells accumulate in tissues bearing lung14, possibly indicating that pDCs in lung are under inflammatory conditions and can adopt DC restricted to the alveolar airspace and only present in parenchyma characteristics3,5,20,23,25. In vivo data suggests it takes at least during pathological situations. 24–48 h for monocytes to differentiate into DCs39 and in vitro, By mapping alveolar airspace MP subsets relative to blood generation of DCs from monocytes (human) or bone-marrow populations, we were able to compare MP populations across derived cells (mouse) takes 5–7 days40,41. In view of our 8-h time- compartments upon LPS-delivery into the airway. In most pre- course, the rapid accumulation of CD1c+ DCs into alveolar air- vious studies, tissue inflammatory composition has been exam- space following LPS inhalation is most in keeping with recruit- ined in isolation without reference to blood and origins have been ment from blood and is consistent with previous observations of inferred35–37. Due to our referencing of blood it appeared logical CD1c+ DCs accumulating in bronchial (conducting airway) that CD14++CD16− MPs infiltrating BAL would be monocytes mucosa within 4 h of allergen challenge35, though detailed ana- recruited from blood with tissue adaptation, but the nature of lyses of the infiltrating population was not possible in this earlier tissue adaptation required investigation. In both mouse and study. Transcriptome analysis further distinguished in vitro human studies, monocytes recruited to tissues can acquire DC or monocyte-derived DCs from LPS-BAL DCs. Our data suggest macrophage characteristics20,22,26. Monocytes have also been that in acute tissue inflammation, CD1c+ DCs, including a subset described entering tissue with minimal adaptation38. The type of with shared expression of monocyte genes, can be directly MP characteristics adopted is highly relevant to the subsequent recruited from blood and arise independently of monocyte-

8 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

a 3.0 MRC1 BTLA+ CD5 BTLA BTLA– 2.5

S100A8 BST1 ZEB1 2.0 CLEC4E FCGR2A CD14 PLA2G2E CD79A CLEC6A KIT CD24 C19orf59 S100A9 vlaue) LILRB2 TMEM14A DPP4 CD163 p 1.5 CXCL9 LILRA4 CD79B TNFRSF13C BST1 HLA-DOB CXCR3 ETS1 S100A8 CD207 TLR3 CD14 F13A1CD163 FCGR2A/C SLC2A1 MS4A1 1.0 TLR4 S100A9 –Log10( HLA-

DC2/3 defining genes DOB 0.5 HLA- DPB1 0 0 5 10 15 Log2(gene expression) –4 –2 024 6 Log2(fold change BTLA+/BTLA–)

b Gated on CD1c+ cells ns HC blood SS-BAL LPS-BAL 1.5 ** 105 105 105 104 104 104 1.0 45% 16% 31% 103 103 103 53% 80% 69% 0.5

0 0 0 BTLA+/– ratio BTLA

3 4 5 3 4 5 3 4 5 0 010 10 10 010 10 10 0 10 10 10 Blood SS LPS CD11c

c d Gated on viable cells e AM CD14++CD16– Gated on DC5 5 5 10 Relative 10 105 SS-BAL 4 0 5 10gene 104 104 10 expression 103 103 103 0 0 0 2 2 3 4 3 4 5 5 0 0 2 3 4 5 0 10 10 10 10 10 10 10 10 10 10 10 10 DC2 DC3 105 105 105 LPS-BAL 4 HLA-DOB 104 104 10 BTLA 3 IL18R1 IL-17 103 103 10

CD5 0 0 CD123 SLAMF7 0 ZEB1 2 2 3 4 3 4 5 5 TLR3 0 0 2 3 4 5 0 10 10 10 10 10 10 10 KIT 10 10 10 10 CXCR3 IFNγ 10 DPP4 CD11c SLC2A1 100 50 IL18RAP ** FCGR3A/B 80 40 TNFRSF9 CASP10 γ 60 30 CXCL9

MARCO IFN 40 20 CFD CLEC4E 20 10 CSF1R % CD11c+ DC5 % T cells producing BST1 0 0 TLR4 AM 14 DC2 DC3 MRC1 30 CLEC7A CD14 CD163 TLR2 20 S100A8

S100A9 IL-17 10 BAL BAL Blood Blood

DC3 DC2 DC2 DC3 % T cells producing 0 AM 14 DC2 DC3

differentiation. This finding is likely to reflect the early time-point We propose that resident AMs operate to recruit DCs into the we were able to study in our experimental model. Later char- alveolar airspace, in addition to their role in recruiting neu- acterization of BAL following experimental inflammatory stimuli trophils and monocytes. Activation of AMs by LPS resulted in and analysis of chronic disease states will reveal whether rapid upregulation of chemokines involved in leukocyte migra- monocyte-derived DCs arise subsequent to this initial inflam- tion and pro-inflammatory cytokine mediators, elucidating how matory DC influx. robust neutrophil recruitment and lung inflammation can be

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

Fig. 4 The pool of recruited LPS-BAL CD1c+ DCs contains DC2 and DC3, which are functionally altered on entry to the inflamed airspace. a Volcano plot shows DEGs between BTLA+ and BTLA− CD1c+ DCs from HC blood. DEGs were calculated by unpaired t-test with p-value adjustment using Benjamini- Yekutieli method. Genes with p < 0.05 are displayed as text. Accompanying scatterplots verifiy that that blood BTLA+ and BTLA– CD1c+ DCs exhibit the expected gene expression profile of DC2 and DC3 respectively. Plots show expression from n = 3 samples with bars showing mean ± SEM. b Comparison of BTLA+:BTLA− ratio in CD1c+ DCs from HC blood, SS-BAL and LPS-BAL by flow cytometry. The gating strategy represented in Fig. 1b was used to define CD1c+ cells. Flow cytometry plots show percentage of BTLA+ and BTLA− CD1c+ cells in representative examples of blood, SS-BAL and LPS-BAL. Accompanying bar graph summarizes flow cytometry data from n = 4–9 replicates. Statistical comparison by one-way ANOVA with Dunnett’s multiple comparisons tests. ns p > 0.05, **p < 0.01. c Heatmap showing expression of the 29 genes with differential expression between blood DC2 and DC3 (p < 0.05 and fold difference >2) in DC2 and DC3 sorted from LPS-BAL and HC blood. Samples were clustered using Euclidian distance metric. d Flow cytometry read-out of a T cell cytokine production assay following 10-day co-culture of allogeneic peripheral blood CD4+ T cells with macrophages and DCs isolated from LPS-BAL. Accompanying bar graphs summarize flow cytometry data showing mean ± SEM IFN-γ and IL-17 production following co-culture of SS-BAL (blue bars) or LPS-BAL (red bars) macrophages and DCs (each n = 2). Two-way ANOVA of IFN-γ production gives ***p < 0.001 for SS-BAL vs. LPS-BAL and ns p = 0.07 for MP type. Two-way ANOVA of IL-17 production gives ns p = 0.06 for SS-BAL vs. LPS-BAL and ns p = 0.67 for MP type. e Flow cytometry plots show representative examples of CD123 and CD11c expression by Axl+Siglec6+ DC5 in SS-BAL and LPS-BAL. The gating strategy represented in Fig. 1b was used to define DC5. Gates demarcate CD11c-expressing cells. Accompanying scatterplot summarizes flow cytometry data, showing proportions of DC5 expressing CD11c in SS-BAL (blue circles) and LPS-BAL (red squares). **p < 0.01 by unpaired t-test observed even after monocyte-depletion in a human LPS inha- infections; current respiratory tract infection; taking prescription medication lation model42. The acute inflammatory cascade initiated by (except oral contraceptives); current smoking or history of smoking 20 cigarettes resident AMs is further amplified by recruited CD14++CD16− per day for more than 2 years or any smoking in the past year; alcohol intake of more than 21 units per week; pregnancy or lactation; abnormal examination monocyte-macrophages. findings at screening (temperature >37.3 °C or oxygen saturation <95% breathing There are a number of limitations to this study. LPS inhalation room air); abnormal blood results at screening (hemoglobin concentration, total provides the opportunity to study early time points in inflam- white cell count or neutrophil count outside the gender-specific laboratory refer- 9 −1 mation but it is not an accurate representation of human disease. ence ranges; platelet count <100 ×10 l ; serum sodium, potassium, creatinine or alanine aminotransferase outside laboratory reference ranges; serum urea >10 mg Infective stimuli are rarely presented to the immune system as dl−1; serum bilirubin >30 μmol l−1); abnormal spirometry at screening (FEV1 or single bolus and will frequently be accompanied by other FVC <80% predicted or FEV1:FVC ratio <70%). Eligible volunteers were given pathogen-associated molecular patterns. BAL permits effective verbal and written information about the study and at least 24 h to consider their sampling of the alveolar space, but “leaves behind” alveolar epi- participation before signing a consent form. All primary research documents were fi anonymised and participant details kept confidentially in accordance with Caldi- thelial cells, adherent in ltrating cells, and those migrating cott guidelines. Data from 19 participants are presented (n = 10 LPS, n = 9 saline). through the interstitium. These populations could make impor- tant contributions to the regulation of the inflammatory response but cannot be sampled without recourse to biopsies that are Study interventions. Volunteers were allocated to inhale 54 μl sterile 0.9% sodium chloride with or without 60 μg LPS from E. coli 026:B6 (Sigma). Participants were unlikely to be ethical in healthy volunteers and may be subject to allocated sequentially to receive saline or LPS to give optimal control over sampling error. Furthermore, BAL itself is inflammatory to the downstream experiments and they were not made aware of which intervention they alveolar region, limiting the opportunity to study serial cell had received. Delivery was targeted at the lower airways using an automatic migration in the same individual, or to assess “recovery” BAL inhalation-synchronized dosimeter nebulizer (Spira, Hameenlinna, Finland). The test solution was released following inhalation of 50 ml air to ensure that laminar samples in this setting. flow was established. Participants performed a 5 s breath hold at vital capacity to Our study has permitted detailed phenotypic, transcriptional, promote deposition of LPS in the lower respiratory tract. and functional analyses of MPs present in the alveolar airspace of Participants were asked about symptoms (flu-like symptoms, sore throat, healthy volunteers inhaling saline control or LPS to induce acute cough, wheeze, chest pain, sputum production, nasal secretions, or any other fl symptom) immediately after inhalation and at 6 h. Body temperature was local in ammation. In LPS-BAL, we observed perturbation of the measured hourly until 6 h. Venous blood samples were obtained at 0, 2, 4, 6, and ++ − MP profile, noting significant influx of CD14 CD16 24 h after inhalation. We used blood leukocyte counts as the indicator of LPS effect. monocyte–macrophages and selected DC subsets, likely regulated Flexible fiber-optic bronchoscopy was performed 8 h after inhalation. by resident AMs. This finding adds to our understanding of the Intravenous sedation with midazolam was available but all participants elected for a non-sedated procedure. Participants received topical administration of 1% potential role of blood DCs and the possibility of monocyte- lignocaine spray to the mouth and pharynx. Bronchial wash of the upper airways independent inflammatory DCs at early time points in acute was performed with 20 ml warmed 0.9% sodium chloride and discarded. BAL of inflammation. Understanding the in vivo kinetics and dynamic the right middle lobe was performed with 150 ml warmed 0.9% sodium chloride regulation of MP following local LPS stimulation extends biolo- and retrieved by gentle suction. gical insights into the mechanisms of acute inflammation. Dis- section of the functional consequences for the host will provide Participant safety. History, bedside observations, cardio-respiratory examination, the opportunity to understand both beneficial and detrimental and spirometry were performed immediately before LPS or saline inhalation. effects of such inflammation. Bedside observations were repeated hourly and spirometry repeated at 6 h after inhalation. If FEV1 fell by 10% from baseline, bronchoscopy was canceled. Prior to bronchoscopy, participants were fasted for 4 h. There was constant monitoring of Methods oxygen saturations and electrocardiogram during bronchoscopy. Patients were Ethical approval. The LPS inhalation study was approved by Newcastle & North observed for 30–60 min after bronchoscopy and allowed to leave if bedside Tyneside 2 Research Ethics Committee of the NHS Health Research Authority observations and cardio-respiratory examinations were normal. Written and verbal (REC reference number 12/NE/0196) under the full title “A lipopolysaccharide advice was given to avoid eating and drinking within two hours of local anesthetic (LPS) inhalation model to characterize divergent cellular innate immune responses administered to the mouth and pharynx. Participants were informed that LPS and presence of alveolar leak early in the course of acute lung inflammation”. The inhalation and bronchoscopy may result in temperature, mild headache, shivering, study was conducted according to the principles expressed in the Helsinki dry cough, and upper airway discomfort. Declaration and informed consent was obtained from all participants. Cell isolation. Blood was collected into EDTA. PBMC were isolated by density Study population. The study was advertised to university undergraduates. centrifugation using Lymphoprep (Stemcell technologies) according to manu- Potential recruits were invited to a screening interview to assess their suitability for facturers instructions. BAL samples were kept at room temperature for 60–90 min participation. Inclusion criteria were healthy adult volunteers aged 18 to 40 able to before processing at 4 °C. BAL fluid was passed through a 100 µ filter (Falcon). give informed consent. Exclusion criteria were age <18 or >40 years; history of Following centrifugation, cell-free supernatant was stored at −80 °C and cells were chronic respiratory disease, diabetes, heart disease, renal disease or recurrent prepared for immediate analytical flow cytometry and cell sorting.

10 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

ab BAL AM CCL2 CCL3 CCL4 SS LPS 1500 **** 400 ** 1000 **** IL8 800 CXCL10 300 1000 CCL3 600 CCL4 200 GBP1 400 500 IL1RN 100 CCL20 200 CXCL11 ICAM1 0 0 0 TNFAIP6 CCL5 CCL11 CXCL1 CLEC4E 15 15 600 ns CCL23 *** ** TLR2 NFKB2 10 10 400 SLAMF7 CFB CSF1 5 5 200

CXCL1 ng/ml BAL pg/ml BAL RELB CMKLR1 JAK3 0 0 0 POU2F2 CCL8 IL8 CXCL10 CXCL12 200 2000 CCL22 **** **** 1500 **** KCNJ2 EBI3 150 1500 CCR5 1000 TNFRSF4 100 1000 SLAMF1 GZMB 500 500 CTSG 50 CD6 0 Relative 0 0 gene SS-BAL supernatant 0 7 14 expression LPS-BAL supernatant cd CCR2 CCR1 CCR5 LPS-BAL CD14++CD16- 12 Ligand(s) CCL2 12 CCL3, CCL412 CCL4, CCL5 LPS-BAL AM

8 8 8 20

4 4 4

15 0 0 0

CXCR1 CXCR2 12 12 Ligand(s) IL-8 CXCL1, IL-8 10 8 8

4 4 HC blood MPs Classical monocyte 5

Log2 gene expression 0 0 DC3

DC2 Log2 cytokine concentration (pg/ml) CXCR3 CXCR4 pDC 12 Ligand(s) CXCL10 12 CXCL12 0 β IL6 IL8 IL1 TNF 8 8 IL10 IL12-p70 4 4

0 0

Flow cytometry and sorting. Cell pellets were washed in PBS (Sigma) supple- anti-CD20-FITC (L27; BD; 345792; 1:25), anti-CD56−FITC (NCAM16; BD; mented with 2% fetal calf serum (Sera lab) and 2 mM EDTA (Invitrogen). The 345811; 1:25), anti-HLA-DR-BV786 (G46−6; BD; 307642; 1:25); anti-CD14− following antibodies were applied for 20 min at 4 °C (clone; supplier; catalog BV650 (M5E2; Biolegend; 301835: 1:40), anti-CD16-PE/Dazzle-594 (3G8; Biole- number; final concentration): anti-CD45-V500 (HI30; BD; 560777; 1:25), anti- gend; 302054; 1:33), anti-CD11c-APCCy7 (Bu15; Biolegend; 337218; 1:33), anti- CD3-FITC (SK7; BD; 345763; 1:25), anti-CD19-FITC (AG7; BD; 347543; 1:25), CD123-PERCPCy5.5 (7G3; BD; 558714; 1:25), anti-CD1c-PECy7 (L161; Biolegend;

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4

Fig. 5 Alveolar macrophages recruit monocytes and DCs and secrete pro-inflammatory cytokines upon acute LPS challenge in vivo. a Heatmap of immune genes showing differential expression between AM isolated from saline BAL and LPS BAL. Genes with>5 fold difference in expression and p < 0.05 are shown. Genes discussed in the text are colored burgundy. b Quantification of chemokines in SS-BAL (blue bars) and LPS-BAL (red bars) supernatant by multiplexed ELISA (Luminex). Bars show mean ± SEM. By unpaired t-test, ns, not significant, **p < 0.01, ***p < 0.001, ****p < 0.0001. c Chemokine receptor gene expression on HC blood classical monocytes (green), DC2 (yellow), DC3 (pink) and pDC (cyan) quantified by Nanostring. Bars show mean ± SEM of n = 3 for each subset. Log2 gene expression is shown. Ligands for each receptor are listed in gray italic type. d Inflammatory cytokine production by CD14++CD16− cells and AM retrieved from LPS BAL (each n = 4) and stimulated ex-vivo with LPS 100 ng ml−1 for 10 h. Cytokines were measured by cytometric bead array. Bars show mean ± SEM. CD14++CD16− cells were compared with AM by t-tests with Holm-Sidak multiple comparison correction and alpha < 0.05. No comparisons were significant

331506; 1:40), anti-CD11b-APC (ICRF44; Biolegend; 301309; 1:25), anti-BTLA-PE cells were stained with a Zombie Aqua (Biolegend) prior to fixation and per- (J168–540; BD; 558485; 1:100), anti-Axl-APC (108724; R&D Systems; FAB154A: meabilization (BD) according to manufacturers instructions. Cells were prepared 1:33) and anti-Siglec-6-A700 (767329; R&D Systems; FAB2859N; 1:33). CD1c+ DC for flow cytometry as above using anti-IFN-γ-PE/Dazzle-594 (4s.B3; Biolegend; phenotyping experiments additionally used anti-CD1a-AF700 (HI149; Biolegend; 505845; 1:100), anti-IL17-AF647 (BL168; Biolegend; 512309; 1:100) and anti-IL4- 300120; 1:50), anti-CD206-PE (19.2; BD; 555954, 1:25), anti-FcER1-PE (CRA1; PECy7 (MP4–25D2; Biolegend; 500823; 1:100) and analyzed on a Fortessa X20 eBioscience; 12–5899–42; 1:25) and anti-CD86-PE (FUN-1; BD; 555658; 1:25). (BD) running FACSDIVA version 7. Dead cells were excluded with DAPI (Partec). Analytical flow cytometry was performed on a BD Fortessa X20 or BD FACSCanto II and cell sorting with a BD Ex-vivo macrophage stimulation. Alveolar macrophages and CD14++CD16− FACS Fusion running FACSDiva version 7. Consistent instrument performance monocyte/macrophages were sorted from LPS BAL (n = 4 each). 1 ×105 sorted was ensured by running Cytometer Setup and Tracking Beads (BD). PBMC were cells were cultured in 96-well round bottom plates in 100 μl RPMI 1640 with periodically run through the analysis template as a biological control to ensure supplements (as above) containing 100 ng ml−1 LPS from E. coli (Sigma). After gates were capturing the expected populations. Doublets were excluded with a SSC- 10 h, supernatants were harvested and stored at −80 °C. Supernatants were batch- H vs. SSC-A plot. FlowJo version 9.6.7 was used for analysis. Sorting strategies are analysed by cytometric bead array (BD) to detect IL-10, IL-1β, TNF, IL-6, and IL-8. given in Supplementary Fig. 3. Bead populations were resolved on a FACS Canto II running FACSDIVA Version 7 and analyzed using FCAP Array software version 3. Peripheral blood MP quantitation. Quantitation of neutrophils and total monocytes was by complete blood count (CBC) in a UKAS accredited clinical Generation of monocyte-derived dendritic cells (moDC). Classical monocytes laboratory using the Sysmex XE-2100. For MP subsets not quantitated in the CBC, were isolated from healthy control peripheral blood by density centrifugation and 1 ml whole blood was lysed using an in-house ammonium chloride lysis buffer and FACS sorting. 5 ×105 monocytes were cultured in 500 μl RPMI 1640 with sup- transferred to a Trucount tube (BD). Antibodies were applied as above and flow fi plements (as above) in 24-well plates for 5 days at 37 °C and 5% CO2. Medium cytometry performed following a nal red cell lysis step. Data were acquired on a contained 50 ng ml−1 GM-CSF (R&D) and 50 ng ml−1 IL-4 (Immunotools). BD Fortessa and cell concentrations calculated with reference to bead counts. Medium and cytokines were refreshed on day 3. Cells were harvested and FACS- sorted on day 5, to exclude undifferentiated CD14+ monocytes and include only Chemokine/cytokine analysis. Chemokines and cytokines in BAL fluid super- CD1c+ moDC. natant were quantified using multiplexed ELISA (ProcartaPlexTM 34−plex, EBioscience). Captured analytes were detected on a Qiagen Liquichip 200 running Statistical analysis. Statistical analyses stated in the text were performed using Luminex 100 integrated system software version 2.3. Procartaplex Analyst version fi GraphPad Prism version 7.0. Gene expression analyses were performed in Multi- 1.0 was used to de ne standard curves and determine analyte concentrations. Experiment Viewer version 10.2 and nSolver v4 advanced analysis module.

Gene expression analysis. Lysates from 1–2×104 cells per sorted population were prepared using 5μl buffer RLT (Qiagen) with 1% beta-mercaptoethanol Data availability (Sigma). Transcripts from 579 immunology genes were detected using a Nano- NanoString gene expression data have been deposited in Gene Expression Omnibus fi String Immunology v2 panel according to the manufacturers instructions. (GEO) with the accession code GSE126923. Other data that support the ndings of this Experiments using blood and BAL DC2/3 included an additional 30-gene add-on study are available from the corresponding author upon reasonable request. to the Immunology v2 panel, designed to detect DC and monocyte/macrophage genes, including ASIP, DBN1, MERTK, C19orf59, F13A1, NDRG2, CCL17, FGD6, Received: 19 July 2018 Accepted: 2 April 2019 PACSIN1, CD1C, FLT3, PPM1N, CD207, GCSAM, PRAM1, CLEC10A, GGT5, S100A12, CLEC9A, Ki67, SIRPA, CLNK, LPAR2, TMEM14A, COBLL1, LYVE1, UPK3A, CXCL5, MAFF, ZBTB46. Panels included 15 housekeeping genes. Data normalization was performed in nSolver version 3, including a background sub- traction step, positive control normalization using synthetic spike-in controls and content normalization using a combination of 15 housekeeping genes. References 1. Scott, C. L., Henri, S. & Guilliams, M. Mononuclear phagocytes of the T cell proliferation assay. T cells were isolated by negative selection from healthy intestine, the skin, and the lung. Immunol. Rev. 262,9–24 (2014). control peripheral blood collected into EDTA using RosetteSep Human T Cell 2. Ginhoux, F. & Guilliams, M. Tissue-resident macrophage ontogeny and Enrichment Cocktail and Lymphoprep (Stemcell technologies). Aliquots were homeostasis. Immunity 44, 439–449 (2016). cryopreserved at −80 °C. Thawed T cells were labeled with 1 μM CSFE (Invitrogen) 3. Tamoutounour, S. et al. Origins and functional specialization of macrophages prior to co-culture. MP populations were sorted from BAL and co-cultured with T cells at a 1:25 ratio in 96 well v-bottom plates containing RPMI 1640 (Lonza) and of conventional and monocyte-derived dendritic cells in mouse skin. − − Immunity 39, 925–938 (2013). with 100 U ml 1 penicillin, 10 μg ml 1 streptomycin, 2mM L-glutamine (all Invi- trogen) and 10% heat-inactivated fetal calf serum (Sera Lab). Cultures were 4. Varol, C. et al. Monocytes give rise to mucosal, but not splenic, conventional maintained for 7 days at 37 °C and 5% CO Outputs were prepared for flow dendritic cells. J. Exp. Med. 204, 171–180 (2007). 2. fl cytometry as described above using anti-CD3-V500 (UCHT1; BD; 561416; 1:50), 5. Zigmond, E. et al. Ly6C hi monocytes in the in amed colon give rise to fl anti-CD4-PE (RTPA-T4; BD; 555347; 1:50), anti-CD8-APCCy7 (SK1; BD; 557834; proin ammatory effector cells and migratory antigen-presenting cells. 1:50) and analzyed on a FACS Canto II running FACSDIVA version 7. Three Immunity 37, 1076–1090 (2012). experiments were performed on populations sorted from LPS BAL. 6. Haniffa, M., Bigley, V. & Collin, M. Human mononuclear phagocyte system reunited. Semin. Cell Dev. Biol. 41, 59–69 (2015). 7. Bachem, A. et al. Superior antigen cross-presentation and XCR1 expression T cell cytokine production assay . CD4 T cells were isolated by negative selection define human CD11c+CD141+cells as homologues of mouse CD8+dendritic from healthy control peripheral blood collected into EDTA using RosetteSep cells. J. Exp. Med. 207, 1273–1281 (2010). Human CD4 T cell enrichment cocktail. Aliquots were cryopreserved at −80 °C. 8. Crozat, K. et al. The XC chemokine receptor 1 is a conserved selective MP populations were sorted from SS-BAL or LPS-BAL and co-cultured with T cells marker of mammalian cells homologous to mouse CD8alpha+dendritic cells. at a 33:1 ratio in 96-well round-bottom plates containing RF10. On day 6, medium J. Exp. Med. 207, 1283–1292 (2010). was replenished with RPMI 1640 with supplements as above and 10 U ml−1 9. Poulin, L. F. et al. Characterization of human DNGR-1+BDCA3+leukocytes human recombinant IL-2 (Immunotools). On day 10, cells were stimulated with 10 ng ml−1 phorbol 12-myristate 13-acetate (Sigma) and 1 μg ml−1 ionomycin as putative equivalents of mouse CD8alpha+dendritic cells. J. Exp. Med. 207, (Sigma) for 4 h, with 10 with brefeldin A (Sigma) added after the first hour. Dead 1261–1271 (2010).

12 NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09913-4 ARTICLE

10. Jongbloed, S. L. et al. Human CD141+(BDCA-3)+dendritic cells (DCs) 36. Lowes, M. A. et al. Increase in TNF-alpha and inducible nitric oxide synthase- represent a unique myeloid DC subset that cross-presents necrotic cell expressing dendritic cells in psoriasis and reduction with efalizumab (anti- antigens. J. Exp. Med. 207, 1247–1260 (2010). CD11a). Proc. Natl Acad. Sci. USA 102, 19057–19062 (2005). 11. Haniffa, M. et al. Human tissues contain CD141hi cross-presenting dendritic 37. Brittan, M. et al. A novel subpopulation of monocyte-like cells in the human cells with functional homology to mouse CD103+nonlymphoid dendritic lung after lipopolysaccharide inhalation. Eur. Respir. J. 40, 206–214 (2012). cells. Immunity 37, 60–73 (2012). 38. Jakubzick, C. et al. Minimal differentiation of classical monocytes as they 12. Merad, M., Sathe, P., Helft, J., Miller, J. & Mortha, A. The dendritic cell survey steady-state tissues and transport antigen to lymph nodes. Immunity lineage: ontogeny and function of dendritic cells and their subsets in 39, 599–610 (2013). the steady state and the inflamed setting. Annu. Rev. Immunol. 31, 563–604 39. Sen, D. et al. Tracking the spatial and functional gradient of monocyte-to- (2013). macrophage differentiation in inflamed lung. PLoS ONE 11, e0165064 (2016). 13. Reynolds, G. & Haniffa, M. Human and mouse mononuclear phagocyte 40. Sallusto, F. & Lanzavecchia, A. Efficient presentation of soluble antigen by networks: a tale of two species? Front. Immunol. 6, 330 (2015). cultured human dendritic cells is maintained by granulocyte/macrophage 14. Desch, A. N. et al. Flow cytometric analysis of mononuclear phagocytes in colony-stimulating factor plus interleukin 4 and downregulated by tumor nondiseased human lung and lung-draining lymph nodes. Am. J. Respir. Crit. necrosis factor alpha. J. Exp. Med. 179, 1109–1118 (1994). Care Med. 193, 614–626 (2016). 41. Inaba, K. et al. Generation of large numbers of dendritic cells from mouse 15. Yu, Y.-R. A. et al. Flow cytometric analysis of myeloid cells in human blood, bone marrow cultures supplemented with granulocyte/macrophage colony- bronchoalveolar lavage, and lung tissues. Am. J. Respir. Cell Mol. Biol. 54, stimulating factor. J. Exp. Med. 176, 1693–1702 (1992). 13–24 (2016). 42. Barr, L. C. et al. A randomized controlled trial of peripheral blood 16. Bharat, A. et al. Flow cytometry reveals similarities between lung macrophages mononuclear cell depletion in experimental human lung inflammation. in humans and mice. Am. J. Respir. Cell Mol. Biol. 54, 147–149 (2016). Am. J. Respir. Crit. Care Med. 188, 449–455 (2013). 17. Baharom, F. et al. Dendritic cells and monocytes with distinct inflammatory responses reside in lung mucosa of healthy humans. J. Immunol. Baltim. Md 1950 196, 4498–4509 (2016). Acknowledgements 18. Patel, V. I. et al. Transcriptional classification and functional characterization This work was funded by Wellcome Trust Clinical Research Training Fellowship of human airway macrophage and dendritic cell subsets. J. Immunol. Baltim. WT097941 (L.J.) with additional contribution from the Newcastle upon Tyne Hospitals Md 1950 198, 1183–1201 (2017). Healthcare Charity and the NIHR Newcastle Biomedical Research Center (A.J.S.). We 19. Villani, A.-C. et al. Single-cell RNA-seq reveals new types of human blood would like to thank the Clinical Research Facility of the Newcastle upon Tyne Hospitals dendritic cells, monocytes, and progenitors. Science 356, eaah4573 (2017). NHS Foundation Trust and Newcastle University and the Endoscopy Unit at the 20. León, B., López-Bravo, M. & Ardavín, C. Monocyte-derived dendritic cells Royal Victoria Infirmary, Newcastle. We appreciate and acknowledge support from formed at the infection site control the induction of protective T helper 1 the Newcastle University NanoString Service and Dr Venetia Bigley, and the Newcastle responses against Leishmania. Immunity 26, 519–531 (2007). University Flow Cytometry Core Facility. 21. Nakano, H. et al. Blood-derived inflammatory dendritic cells in lymph nodes stimulate acute T helper type 1 immune responses. Nat. Immunol. 10, Author contributions 394–402 (2009). L.J., S.W., G.R., D.M., A.Fu, K.G., A.F., I.F., M.H.R.S., J.S., M.C., M.H., and A.J.S. 22. Bain, C. C. et al. Resident and pro-inflammatory macrophages in the colon Conceptualization: A.J.S., M.H., M.C., L.J. Methodology: A.J.S., S.W., D.M., A.Fu, represent alternative context-dependent fates of the same Ly6Chi monocyte K.G., A.F., M.C., M.H., A.J.S., L.J. Investigation: S.W., L.J., G.R., I.F., M.H.R.S., J.S. precursors. Mucosal Immunol. 6, 498–510 (2013). Writing-original draft: L.J., A.J.S., M.H. Writing-revision and editing: LJ, AJS, MH 23. Plantinga, M. et al. Conventional and monocyte-derived CD11b(+) dendritic Funding acquisition: A.J.S., L.J. Resources: A.J.S., M.H., M.C. Supervision: A.J.S., cells initiate and maintain T helper 2 cell-mediated immunity to house dust M.H., M.C. mite allergen. Immunity 38, 322–335 (2013). 24. Iijima, N., Mattei, L. M. & Iwasaki, A. Recruited inflammatory monocytes stimulate antiviral Th1 immunity in infected tissue. Proc. Natl Acad. Sci. USA Additional information 108, 284–289 (2011). Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- 25. Segura, E. & Amigorena, S. Inflammatory dendritic cells in mice and humans. 019-09913-4. Trends Immunol. 34, 440–445 (2013). 26. Segura, E. et al. Human inflammatory dendritic cells induce Th17 cell Competing interests: The authors declare no competing interests. differentiation. Immunity 38, 336–348 (2013). 27. Martinez, F. O., Gordon, S., Locati, M. & Mantovani, A. Transcriptional Reprints and permission information is available online at http://npg.nature.com/ profiling of the human monocyte-to-macrophage differentiation and reprintsandpermissions/ polarization: new molecules and patterns of gene expression. J. Immunol. Baltim. Md 1950 177, 7303–7311 (2006). Journal peer review information: Nature Communications thanks Stefan Zielen, 28. McGovern, N. et al. Human dermal CD14+ cells are a transient population of Michael Matthay and the other anonymous reviewer(s) for their contribution to the peer monocyte-derived macrophages. Immunity 41, 465–477 (2014). review of this work. Peer reviewer reports are available. 29. Jardine, L. et al. Rapid detection of dendritic cell and monocyte disorders using CD4 as a lineage marker of the human peripheral blood antigen- Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in presenting cell compartment. Front. Immunol. 4, 495 (2013). published maps and institutional affiliations. 30. Yin, X. et al. Human blood CD1c+dendritic cells encompass CD5 high and CD5 low subsets that differ significantly in phenotype, gene expression, and functions. J. Immunol. Baltim. Md 1950 198, 1553–1564 (2017). Open Access This article is licensed under a Creative Commons 31. Schlitzer, A. et al. IRF4 transcription factor-dependent CD11b+dendritic cells Attribution 4.0 International License, which permits use, sharing, in human and mouse control mucosal IL-17 cytokine responses. Immunity 38, adaptation, distribution and reproduction in any medium or format, as long as you give 970–983 (2013). appropriate credit to the original author(s) and the source, provide a link to the Creative 32. Granot, T. et al. Dendritic cells display subset and tissue-specific maturation Commons license, and indicate if changes were made. The images or other third party dynamics over human life. Immunity 46, 504–515 (2017). material in this article are included in the article’s Creative Commons license, unless 33. Korenfeld, D. et al. A type of human skin dendritic cell marked by CD5 is indicated otherwise in a credit line to the material. If material is not included in the associated with the development of inflammatory skin disease. JCI Insight 2, ’ 96101 (2017). article s Creative Commons license and your intended use is not permitted by statutory 34. Dijkstra, D. et al. Identification and quantification of basophils in the airways regulation or exceeds the permitted use, you will need to obtain permission directly from of asthmatics following segmental allergen challenge. Cytom. Part A 85, the copyright holder. To view a copy of this license, visit http://creativecommons.org/ 580–587 (2014). licenses/by/4.0/. 35. Jahnsen, F. L. et al. Rapid dendritic cell recruitment to the bronchial mucosa of patients with atopic asthma in response to local allergen challenge. Thorax © The Author(s) 2019 56, 823–826 (2001).

NATURE COMMUNICATIONS | (2019)10:1999 | https://doi.org/10.1038/s41467-019-09913-4 | www.nature.com/naturecommunications 13